In many industrial automation applications, it is desired to start and stop AC electric motors according to a specified speed profile. Motor control systems have been developed and employed to control speed and other aspects of motor performance during startup using variable frequency drives. Alternatively, induction motor systems often include soft-starters to energize the stator windings during starting and stopping, with the soft startup controls being bypassed once the motor reaches the normal operating speed. To accurately control the motor speed during starting and stopping, it is necessary to measure or estimate the actual rotational speed of the rotor. Many motor control systems employ some form of tachometer or other sensor device mechanically coupled to the motor shaft to produce a feedback signal representing the motor speed, which can then be used in feedback control of the motor during starting, stopping, and steady-state operation. However, such external sensors add cost and weight and require maintenance. Therefore, sensorless systems have been proposed which utilize speed estimation apparatus to derive a speed estimate based on measured electrical motor signals. In certain proposed sensorless speed estimation approaches, the zero-crossing times of the stator winding currents are measured and the corresponding phase angle errors between consecutive zero-crossings are analyzed to produce a signal frequency which is ideally related to the rotor speed. However, experimentation has shown that the detected signal is only strong enough for reliable speed identification over a small portion of the speed range, and the measured signal is easily corrupted by power supply disturbances. Consequently, there is a need for improved motor control apparatus and sensorless speed estimation techniques and systems for motor control applications, particularly for controlling motor speeds during starting and stopping.